248 ALASKAN GLACIER STUDIES
isolated knolls already mentioned and upon one strip of elevated outwash terrace which, has been immune from stream, encroachment since before 1898, and upon which shrubs grew at that time, as photographs show. East of this terrace there is no part of the out-wash plain, except the isolated gravel knobs, that has any shrubs whatever. That streams flow over every part of this surface sometime during practically every year, is indicated by the fact that not even annual plants can secure a permanent foothold. One small strip near the center of the plain that had scattered shrubs in 1898 has since had them destroyed by the streams.
The above conditions apply to the mile or so of outwash plain nearest the glacier, where there are a very few small areas of scattered shrubs, but the next mile and a half or two miles, the width varying in different parts of the plain, has absolutely no vegetation except a few annual plants. The mile, or thereabouts, of plain nearest the fiord, on the other hand, has a predominance of vegetation cover, chiefly cottonwoods and alder, with narrow stream-channel strips extending through the forest. There are trees several feet in diameter, indicating a long period of immunity from extensive and continuous stream encroachment in this part of the outwash plain. This period of immunity is apparently now at an end, for there are lanes of stream channel extending through the forest, in some of which the trees have been removed while in others the dead trees, 40 to 50 feet high, still stand upright. The northern margin of this forest is made up of long points of tree growth, extending up the plain toward the glacier. That this finger-like inner margin is a condition due to the destruction of forest rather than to advance of trees is suggested by the recent killing of mature cottonwoods which had grown to a diameter of two or three feet far up on one of the points. That all the forest has not been removed by the streams may be due to lack of time, the streams possibly now being more active again after a period when the glacier was farther back than now. Or it may be due to the fact that the streams have become entrenched in the gravels at other points as far as this down the outwash plain and while they persist there, vegetation thrives elsewhere. Sometimes the streams do leave these channels and mow down a section of forest, as during the floods in 1905 when trees were uprooted or smothered, and even houses washed away in the town of Valdez. The forest advances farthest toward the glacier in the eastern side of the outwash plain where the largest stream has flowed ever since 1898, and where now many of the stream channels extend down through the forest. We are at a loss to account for this peculiar distribution of tree growth on the Valdez Glacier outwash gravel plain, though it is possible that the point of outflow of the largest glacial stream varies from time to time, the latest point, since 1898, being on the eastern side and the stream from it still being engaged in raising the depression of Robe Lake to the level of the previously-deposited central part of the plain.
Former Advance. The recency of retreat of Valdez Glacier from the latest notable advance is proved by the distribution of vegetation, the clearness of grooving on the valley walls, and the streams cascading down the mountain sides as yet little entrenched in gorges. The possibility of retreat of the glacier from this former stand to a point even farther than at present is suggested by the vegetation-covered talus slopes on the west side of the valley, of which the ones nearest the glacier began to be undercut by glacier and streams before 1898, so that part of the vegetation has been removed.
At the maximum of glaciation the main ice current of Valdez Glacier seems to have been forced over to the west wall by the thrust of the Corbin and two other tributaries from